takazudo/light-review
skills/light-review/SKILL.md
Lightweight code review. Dispatches to GitHub Copilot CLI (/gcoc-review) by default, or to Claude / Codex / Copilot-Opus depending on flags. Use when: (1) Quick review of a small change, (2) Child agents self-reviewing before reporting to manager, (3) User says 'light review' or 'quick review', (4) Review is needed but /deep-review is overkill. Always operates in PR/diff mode.
$ install --global
skillsauthnpx skillsauth add takazudo/claude-resources light-reviewInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 29, 2026, 4:44 AM171.9s1 file scanned
SKILL.md
- name:
- light-review
- description:
- Lightweight code review. Dispatches to GitHub Copilot CLI (/gcoc-review) by default, or to Claude / Codex / Copilot-Opus depending on flags. Use when: (1) Quick review of a small change, (2) Child agents self-reviewing before reporting to manager, (3) User says 'light review' or 'quick review', (4) Review is needed but /deep-review is overkill. Always operates in PR/diff mode.
- argument-hint:
- [-haiku|-so|-op] [-co|-gco|-gcoc]
Light Review
Lightweight code review. Runs whichever reviewers are specified by flags; falls back to the skill's defaults when none are passed.
Review Focus
- Silly mistakes, bugs, and logic errors
- Missing error handling
- Code quality and readability
- Obvious refactoring opportunities
Flags
Model flags (pick at most one — sets Claude model for Claude-based reviewers)
-haiku/--haiku— Claude Haiku-so/--sonnet— Claude Sonnet-op/--opus— Claude Opus
If none passed and no backend flag is passed either, the skill falls to the backend default (-gcoc) — no Claude reviewers run.
If a model flag IS passed, it turns on the Claude-reviewers branch (2 code-reviewer subagents at that model).
If multiple model flags are passed, the last one wins.
Backend flags (combinable — external review tools)
-co/--codex— OpenAI Codex CLI (/codex-review)-gco/--github-copilot— GitHub Copilot CLI, Opus (/gco-review)-gcoc/--github-copilot-cheap— GitHub Copilot CLI, GPT-4.1 free (/gcoc-review)
Multiple backend flags may be combined — each specified backend runs in parallel and findings are consolidated.
Default for this skill: -gcoc (used when neither a model flag nor any backend flag is passed).
Flag-resolution summary
| Flags passed | What runs |
|---|---|
| (none) | /gcoc-review only |
| -op (or -so, -haiku) | 2 Claude reviewers at that model |
| -gco | /gco-review only |
| -co | /codex-review only |
| -op -gco | 2 Opus Claude reviewers and /gco-review in parallel |
| -co -gcoc | /codex-review and /gcoc-review in parallel |
Process
Step 1: Get the Diff
BRANCH=$(git branch --show-current)
BASE=$(gh pr view --json baseRefName -q '.baseRefName' 2>/dev/null)
If no PR exists, use the default branch:
BASE=$(git remote show origin | grep 'HEAD branch' | awk '{print $NF}')
git diff "$BASE"...HEAD
Step 2: Dispatch Reviewers in Parallel
Based on the flags, launch every selected reviewer in the same message (parallel).
Claude branch (only when a model flag is passed)
Launch 2 code-reviewer subagents with model set to haiku / sonnet / opus per the model flag.
Reviewer 1: Bugs & Logic
Review the code changes focusing on:
1. Logic errors, typos, incorrect implementations
2. Missing null checks, off-by-one errors
3. Broken functionality, incorrect API usage
4. Error handling issues
Be concise. Only flag real problems, not style preferences.
REPORTING: Save your FULL findings to the log file (as per your log generation rules).
Then return to the caller ONLY:
- A bullet list of high-priority findings (1 sentence each, max 3 items)
- The log file path
Do NOT return the full analysis — it is in the log file.
Reviewer 2: Quality & Structure
Review the code changes focusing on:
1. Code duplication (DRY violations)
2. Overly complex code that can be simplified
3. Type safety issues
4. Performance concerns (unnecessary re-renders, missing memoization)
5. Better patterns or abstractions
Be concise. Only flag real problems, not style preferences.
REPORTING: Save your FULL findings to the log file (as per your log generation rules).
Then return to the caller ONLY:
- A bullet list of high-priority findings (1 sentence each, max 3 items)
- The log file path
Do NOT return the full analysis — it is in the log file.
Backend branch (only when a backend flag is passed)
For each specified backend, invoke the matching skill in parallel (single message, multiple tool calls):
-co→Skill(skill="codex-review")— silently falls back to Opus (2code-reviewersubagents atmodel: opus) if codex is rate-limited-gco→Skill(skill="gco-review")-gcoc→Skill(skill="gcoc-review")
Each backend skill already handles its own rate-limit / fallback behavior silently. For -co, that fallback is Opus — the user picked -co to mean "the better reviewer," and Opus is the Claude-side stand-in when codex is down.
Default (no flags)
Equivalent to -gcoc. Invoke /gcoc-review only.
CRITICAL: Launch all reviewers (Claude + backend) in parallel in a single message.
Step 3: Synthesize and Apply
After all reviewers complete (each returns high-priority items + log path):
- Merge and deduplicate findings across all reviewers (Claude + backends)
- Categorize by priority (high / medium / low)
- If more detail is needed on a finding, read the reviewer's log file
- Apply high-priority fixes automatically
- Apply medium-priority fixes if clearly safe
- Skip low-priority and style-only suggestions
Step 4: Commit Fixes
If fixes were applied, commit them with a descriptive message.
Important Notes
- This is a lightweight review — keep it fast. The goal is a quick sanity check, not a deep audit.
- Reviewers save full findings to log files, return only high-priority items + path.
- For thorough review (3–6 reviewers), use
/deep-reviewinstead. - Log files are available via
/logreferfor future sessions.
Related Skills
takazudo/dev-linkify-cc-resources
development
VerifiedTrustedCommunity
Link Claude Code skill names mentioned in a CodeGrid article (data/{series}/{n}.md) to the author's public claude-resources repo, pinned to the latest commit hash so links don't rot. Use when: (1) user says 'linkify cc resources', 'link the skills', 'link skill names', or invokes /dev-linkify-cc-resources; (2) editing a CodeGrid article that mentions `/commits`, `/pr-complete`, `/skill-creator` or other Claude Code skills and they should point to claude-resources. Only links skills that actually exist in the public repo; skips hypothetical examples and code blocks.
10SKILL.mdUpdated Jun 2, 2026
takazudo/opus-2nd
development
VerifiedTrustedCommunity
Second opinion from Claude Opus on a plan or approach. Use when: (1) Planning phase of /big-plan needs a higher-quality review than /codex-2nd / /gco-2nd / /gcoc-2nd, (2) User says 'opus 2nd' or 'opus opinion', (3) Wanting Anthropic's larger model to critique a plan. Spawns a general-purpose Agent with model: opus that reads the plan file and returns structured feedback. Anthropic quota — not free.
10SKILL.mdUpdated May 29, 2026
takazudo/dev-ai-based-test
tools
VerifiedTrustedCommunity
AI-based testing via subagent + a per-task test-flow skill. Use when the user wants to verify something that mechanical assertions can't fully capture — image recognition, visual size/position comparison, animation smoothness, multi-step manual flows that need AI judgment. Triggers: 'AI-based test', 'AI test', 'visual verify', 'image recognition test', 'manual operation test', 'human-eye check', 'verify visually', 'compare screenshots', 'looks the same', 'looks correct'. The skill's job is to (1) author a focused test-flow skill that captures the exact procedure + verdict criteria, then (2) dispatch a verification subagent via the Agent tool that loads BOTH the test-flow skill AND a browser-driving skill (/verify-ui primary, /headless-browser fallback) so the subagent has clear context and consistent verdicts. NEVER uses `claude -p` — subagent dispatch goes through the Agent tool exclusively.
10SKILL.mdUpdated May 29, 2026
takazudo/cleanup-resources
development
VerifiedTrustedCommunity
End-of-workflow audit of touched GitHub issues, PRs, and branches via a Sonnet subagent. Use when: (1) /big-plan, /x-as-pr, or /x-wt-teams finishes its main work and needs to verify every touched resource is in the right state (closed when done, kept when ongoing, deleted when dead), (2) User says 'cleanup resources', 'audit cleanup', or 'check what should be closed', (3) A long workflow ends and the manager wants a structured paper trail of what it closed/kept/deleted. Auto-execute by default — the Sonnet agent proposes, the manager (you) executes safe actions and prints a final report.
10SKILL.mdUpdated May 29, 2026